An underground water supply pipeline function cascade evaluation method and electronic equipment

Abstract

The present application relates to the technical field of urban water supply network safe operation and risk assessment, and particularly relates to a kind of underground water supply pipeline function cascade assessment method and electronic equipment.The method comprises: constructing network map model and integrating multi-source monitoring data;Establish the mapping rule from structural failure mode to hydraulic disturbance parameter;Build a knowledge graph that integrates asset attributes and operation and maintenance knowledge, output failure priori and rules;Based on the pressure-driven model, solve the water, quantify the node service capability;Update the structural failure probability by using the monitoring residual and the knowledge priori fusion;According to the hydraulic redistribution result, calculate the pipe segment overload ratio and secondary failure risk;Introduce cascade consistency constraints, apply monotone projection to failure probability sequence, apply monotone regression to system function sequence, eliminate non-physical oscillation;Finally, calculate the functional cascade comprehensive loss, and identify the key weak link and high-risk transmission path through counterfactual analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of safe operation and risk assessment of urban water supply pipe network, and particularly relates to a method for evaluating function cascade of underground water supply pipeline and an electronic device. BACKGROUND

[0002] The urban underground water supply pipe network is a lifeline project for ensuring the operation of social economy and the safety of residents' life. The pipe network is in a complex service environment of coupling of soil corrosion, ground load, water pressure fluctuation and the like for a long time, and is prone to structural failure such as pipe corrosion, interface leakage and pipe body burst. The physical failure of a single pipe section often causes redistribution of the internal hydraulic state (pressure, flow) of the pipe network, resulting in insufficient water supply pressure in a local area or overload of the flow of some pipes. Such hydraulic disturbance may further induce structural secondary failure of adjacent or associated pipe sections, thereby forming a cascade propagation effect of “initial structural failure → hydraulic state disturbance → degradation of water supply service function → induced secondary structural failure”. If this cascade failure process cannot be timely warned and effectively controlled, it may lead to large-scale water supply interruption, causing serious social impact and economic loss.

[0003] Therefore, it is of great significance to accurately evaluate the cascade failure risk of the water supply pipe network, identify the key weak links and high-risk propagation paths in advance, and formulate preventive maintenance strategies, optimize emergency repair schemes and ensure the resilience of the water supply system.

[0004] The existing water supply pipe network risk assessment methods are mainly divided into two categories: model-driven and data-driven. However, in dealing with the complex dynamic process of cascade failure, both have obvious shortcomings:

[0005] (1) Lack of physical consistency constraint: Traditional cascade failure simulation is mostly based on iterative attack or Monte Carlo sampling of hydraulic models. In simulating failure propagation, the phenomenon of “non-physical oscillation” that violates physical common sense often occurs. For example, when the number of failed pipe sections increases, the calculated system service level actually rises, or the failure set appears to be backtracking in the iteration process, resulting in unreliable evaluation results.

[0006] (2) Insufficient knowledge and data fusion: Pure model-driven methods (such as preset attack scenario simulation) cannot effectively fuse real-time monitoring data, and the response to actual faults in progress is lagging. Pure data-driven methods (such as fault diagnosis based on deep learning) highly depend on a large number of high-quality historical fault samples for training. In the scenario where fault samples are scarce and monitoring points are sparsely arranged in the actual pipe network, the model generalization ability is poor, and the decision-making process is like a “black box”, lacking effective use of engineering knowledge such as “valve isolation boundary”, “pipe material brittleness mechanism” and “soil corrosion influence”, resulting in poor interpretability of the evaluation results.

[0007] (3) Structure and function evaluation is fragmented: Existing research often focuses on a single dimension, often only focusing on structural life prediction or only focusing on hydraulic state, lacking a bridging model that quantitatively correlates structural performance degradation and loss of water supply service function, such as service interruption caused by insufficient pressure.

[0008] In summary, the prior art lacks a method that can integrate physical mechanism knowledge and real-time monitoring data into the same evaluation framework with strict cascade process physical consistency constraints, resulting in poor stability, weak interpretability, and low correlation with actual function loss of cascade failure evaluation results. SUMMARY

[0009] The main purpose of the present application is to provide a method for evaluating the function of a cascade of underground water supply pipelines and an electronic device, which aims to solve the above technical problems.

[0010] To achieve the above purpose, on the one hand, the present application provides a method for evaluating the function of a cascade of underground water supply pipelines, comprising the following steps:

[0011] S1, obtaining the basic topological data of the underground water supply network in the evaluation area and constructing a water supply network graph model, obtaining the real-time monitoring hydraulic data of the sensors arranged at the key nodes of the network and mapping them to the established water supply network graph model, providing a basic calculation platform for subsequent hydraulic state solving;

[0012] S2, establishing a structural failure mode library, converting the physical structural failure modes of the pipeline into equivalent hydraulic disturbance parameters;

[0013] S3, constructing a water supply network knowledge graph, inputting the pipe network asset properties and historical operation and maintenance records in the graph model of step S1 as entities, including pipe material, pipe diameter, pipe age, interface form and burial depth, outputting the pipe segment failure prior weight, isolation boundary and secondary failure triggering rule based on the graph reasoning mechanism, as the basis for step S5 prior probability calculation and step S6 cascade determination;

[0014] S4, under the current failure set and valve state, introducing the equivalent hydraulic disturbance parameters generated in step S2, establishing a pipe network hydraulic equation set composed of node continuity equation, pipe segment energy equation, and node pressure and head relationship equation and solving, obtaining node pressure and pipe segment flow, realizing pipe network hydraulic redistribution; and using a pressure-driven water supply model to calculate the node service factor and the actual water supply amount of the node;

[0015] S5, based on the preset monitoring nodes and monitoring pipe segments, obtaining the measured data, constructing the hydraulic residual error as the data evidence, and combining the prior weight output by the knowledge graph in step S3, calculating the posterior probability of structural failure of the candidate pipe segment;

[0016] S6. Calculate the overload ratio of the pipe segment based on the hydraulic redistribution result of the pipe network obtained in step S4. Update the secondary failure probability based on the overload ratio and the secondary failure triggering rule in step S3. Combine the secondary failure probability with the structural failure posterior probability obtained in step S5 to calculate the comprehensive failure probability of the pipe segment. Arrange the comprehensive failure probabilities of the pipe segment at each time step in chronological order to form a comprehensive failure probability sequence.

[0017] S7. Introduce cascaded consistency constraints, perform monotonic projection on the comprehensive failure probability sequence obtained in step S6 to satisfy the monotonically increasing nature of the failure set, and construct a system function index sequence based on the actual water supply of the nodes calculated in step S4. Apply monotonic regression constraints to the system function index sequence to eliminate non-physical function rebound.

[0018] S8. Calculate the functional cascade comprehensive loss index based on the system functional index sequence after consistency constraints in step S7, and quantify the marginal contribution of each pipe segment to the functional cascade comprehensive loss index, thereby identifying key weak links and high-risk cascade paths in the pipeline network.

[0019] Furthermore, step S1 specifically includes:

[0020] S11. Constructing a water supply network diagram model: The expression for the water supply network diagram model is:

[0021] ,

[0022] in: This represents the set of all nodes in the pipeline network. This represents the set of all pipe segments in a pipeline network;

[0023] S12. Obtain monitoring and maintenance data: The monitoring data includes node monitoring pressure. Pipeline section monitoring flow Valve opening / status Pump operating conditions Operation and maintenance data includes work orders and emergency repair records; after unifying the timestamps of the above data and performing missing value completion and anomaly removal, it is assigned to the water supply network diagram model. The corresponding node and pipe segment objects.

[0024] Furthermore, in step S2, the method for converting the physical structural failure mode of the pipeline into equivalent hydraulic disturbance parameters includes:

[0025] S21. For pipe burst or fracture modes, the roughness coefficient of the failed pipe section is... Replace with failure state coefficient Alternatively, the pipe segment can be disconnected within the topology.

[0026] S22. For leakage or perforation modes, introduce a leakage term that varies with pressure at the end node of the failed pipe section. ;

[0027] S23. For local buckling modes, the effective pipe diameter of the pipe section Modified to the reduced failure diameter Or increase the local head loss coefficient;

[0028] The , and All of these are updated as parameters in the network hydraulic equations set described in step S4.

[0029] Furthermore, step S3 specifically includes:

[0030] S31. Construct a knowledge graph of the water supply network, wherein the expression of the knowledge graph of the water supply network is:

[0031] ,

[0032] Among them, entity set This includes pipe sections, nodes, valves, DMA zones, road grades, soil environment, pipe age, material, failure modes, and emergency repair resources; a set of relationships. This includes connectivity relationships, supply relationships, isolation and reachability relationships, same-zone relationships, same-repair team relationships, and failure mode-disturbance type relationships;

[0033] S32. Construct a rule base based on expert knowledge base and historical fault data. By utilizing the association path search of the knowledge graph, the prior weights of the pipe segment failure modes are output. Isolation boundary range and secondary failure triggering rules, where subscripts Indicates the pipe segment index. Indicates the failure mode type.

[0034] Furthermore, step S4 specifically includes:

[0035] S41. Establish the nodal continuity equation, the expression of which is:

[0036]

[0037] in, Indicates the current time step or iteration step of the evaluation; Indicates the current computing node. Represents nodes Adjacent nodes; Let i be the set of adjacent nodes of node i. For the flow sign function, when the flow rate is from Flow direction Add 1 if the time is right, and subtract 1 otherwise. Let k be the flow rate of the pipe segment from node i to node j. Step size Time node The user's water demand, Step size Time node The actual water supply Step size Time node The equivalent leakage amount;

[0038] S42. Establish the pipe segment energy equation based on the Hazen-Williams formula. The expression of the pipe segment energy equation is as follows:

[0039]

[0040] in, Indicate step Time node water head height, Indicate step Time node water head height, Indicates pipe section Length, Indicates pipe section The Heisenberg-Williams coefficient, Indicates pipe section The physical diameter Indicates step size Time Management Section Flow rate; when pipe section At that time, according to the rules of step S2, let or Or disconnect directly to characterize the pipe burst, break, or blockage pattern;

[0041] S43. Establish the equation relating nodal pressure and head. The expression for the equation relating nodal pressure and head is as follows:

[0042]

[0043] in, Let ρ be the nodal pressure, ρ be the water density, and g be the gravitational acceleration. For nodes The altitude;

[0044] S44. The node service factor is calculated using a pressure-driven water supply model, wherein the pressure-driven water supply model is as follows:

[0045]

[0046] wherein, represents a node pressure service factor, is a dimensionless coefficient with a value range of represents a step time node pressure, is a minimum water supply pressure threshold, is a service pressure threshold, is a pressure-water supply nonlinear index;

[0047] S45, based on the node service factor and the water demand, the actual water supply of the node is calculated, and the calculation formula is:

[0048]

[0049] wherein, represents a step time node actual water supply, is a step time node user water demand, represents a node pressure service factor.

[0050] Further, step S5 specifically comprises:

[0051] S51, based on the deviation between the observation value and the simulation value of the monitoring node set and the monitoring pipe section set, a residual error vector is constructed, and the expression of the residual error vector is:

[0052]

[0053] wherein, is a residual error vector, is a monitoring node set, is a monitoring pipe section set; respectively represent the monitoring pressure of node and the pressure at step , respectively represent the monitoring flow of pipe and the flow at step ;

[0054] S52, calculate the correlation between the residual error vector and the preset fault fingerprint, and generate the fault evidence strength combined with the prior weight output by step S3, and the calculation formula of the fault evidence strength is:

[0055]

[0056] wherein, represents a step time pipe section ​the strength of the failure evidence, is a pipe segment the theoretical residual fingerprint when failure occurs, generated by hydraulic simulation, is the prior weight output by the knowledge graph in step S3, , is a weight coefficient; represents a correlation coefficient calculation function;

[0057] S53, construct a pipe segment feature vector containing pipe material, pipe age and burial depth attributes, combine the evidence strength and calculate the structure failure posterior probability through an activation function, the calculation formula of the structure failure posterior probability is:

[0058]

[0059] wherein, represents a step size is the posterior probability of structure failure of the pipe segment at time is a Sigmoid activation function, is a pipe segment feature vector, is an attribute weight vector, is a bias term, is an evidence strength weight coefficient, and the are model parameters trained based on historical failure sample data through a logistic regression or neural network algorithm.

[0060] Further, step S6 specifically includes:

[0061] S61, calculate the pipe segment overload ratio according to the structure-hydraulic comprehensive capacity of the water supply network preset and the pipe segment flow output in step S4, the calculation formula is:

[0062]

[0063] wherein, is the overload ratio, i.e., the structure performance utilization rate, is the structure-hydraulic comprehensive capacity of the pipe segment; represents a step size is the flow of the pipe segment at time

[0064] S62, calculate the secondary failure probability gain based on the overload ratio in step S61, the calculation formula is:

[0065]

[0066] wherein, is the secondary failure probability gain, is an overload sensitivity coefficient;

[0067] and the secondary failure probability gain obtained in this step are fused to calculate the integrated failure probability:

[0068]

[0069] wherein, is the integrated failure probability at the current time step after fusing the data-driven structural failure evidence and the model-driven hydraulic overload risk; if the next time step evaluation is needed, let .

[0070] Further, step S7 specifically comprises:

[0071] S71, monotonically projecting the integrated failure probability sequence obtained in step S6, and the expression of the monotonically projection is:

[0072]

[0073] wherein, wherein, is the final failure decision probability after monotonically correction, is the historical time step index, and the failure set is generated according to the threshold , to ensure that the failure set satisfies ;

[0074] S72, constructing the system function index by using the actual water supply quantity of the node obtained in step S4, and the calculation formula is:

[0075]

[0076] wherein, is the user node set with water demand requirement, is the importance weight of the node ; represents the system function satisfaction rate calculated only based on the current hydraulic state; is the actual water supply quantity of the node at the step length ; is the user water demand of the node at the step length ; represents the node pressure service factor; subsequently, monotonically regression constraint is applied to the same, to obtain the corrected effective system function index , set , and the monotonically regression constraint ensures that the system performance index is monotonically non-increasing with the cascade process;

[0077] ​​​​S73, performing structure-function causal consistency check to ensure that the secondary failure pipe section meets at least one condition of overload trigger, knowledge rule trigger or data evidence trigger:

[0078] When a pipe section is determined to be a secondary failure, i.e. at least one trigger condition must be met, which includes overload trigger, knowledge rule trigger or data evidence trigger;

[0079] wherein the determination condition of the overload trigger is: ;

[0080] The determination condition of the knowledge rule trigger is that the rule base gives a trigger label of isolation inaccessibility or low pressure persistence or material high fragility;

[0081] The determination condition of the data evidence trigger is that the residual fingerprint similarity corresponding to the pipe section exceeds a threshold value.

[0082] Further, step S8 specifically comprises:

[0083] S81, based on the system function index and event weight after consistency constraint in step S7, calculating a functional cascade comprehensive loss index, and the calculation formula of the functional cascade comprehensive loss index is:

[0084]

[0085] wherein, is the functional cascade comprehensive loss index, is the time weight, is the total time step of cascade evaluation, is the system function index after consistency constraint;

[0086] S82, based on the cascade failure under different failure sets, calculating the weak link contribution degree of each pipe section to identify the key weak link and the high-risk cascade path, and the calculation formula of the weak link contribution degree is:

[0087]

[0088] wherein, is the weak link contribution degree, represents the loss obtained by cascade evaluation with the failure set F, is the counterfactual evaluation loss after removing the pipe section from the failure set, and the key weak link is identified in descending order of numerical value.

[0089] ​In a second aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the above-mentioned method for cascade assessment of underground water supply pipeline functions.

[0090] Thanks to the above technical solutions, the present application has the following advantages:

[0091] (1) The consistency constraint is introduced to improve the stability and reliability of the assessment results: the present application applies a monotonic increasing constraint to the failure set to ensure that the pipe sections that have been determined to fail will not appear to be backfitted; and a monotonic non-increasing constraint is applied to the system function index to eliminate the non-physical function rebound phenomenon caused by value fluctuations. These two constraint mechanisms effectively avoid the non-physical oscillation problem in the traditional assessment method, making the assessment results more in line with the physical law of pipe network failure and function degradation, and improving the stability and reliability of the assessment results.

[0092] (2) The knowledge prior and data-driven are fused to expand the applicable scenarios of the method: the present application constructs a water supply network knowledge graph and a rule base, explicitly incorporating engineering knowledge such as pipe material brittleness, isolation boundary, and secondary failure triggering mechanism into the assessment framework; at the same time, the monitoring residual assimilation technology is used to combine real-time monitoring data with knowledge prior to calculate the failure probability of the pipe section. This knowledge-data dual-driven mechanism not only ensures the interpretability of the assessment process, but also improves the applicability of the method in scenarios with sparse monitoring points and insufficient historical samples, solving the limitations of pure model-driven and pure data-driven methods.

[0093] (3) The structure failure and function loss are associated to realize the quantitative assessment of cascade impact: the present application establishes a quantitative correlation model of "structure failure-hydraulic disturbance-function loss", calculates the node service factor and actual water supply amount through the pressure-driven water supply model, and quantifies the impact of cascade failure on the pipe network water supply function using the functional cascade comprehensive loss index; at the same time, by calculating the contribution degree of weak links, the key weak links and high-risk cascade paths with the greatest impact on the pipe network function are identified. This integrated structure and function assessment method provides direct quantitative basis for engineering decision-making such as priority inspection, priority reinforcement, and emergency repair of the pipe network. BRIEF DESCRIPTION OF DRAWINGS

[0094] Figure 1 is a general flowchart of the underground water supply pipeline function cascade assessment method provided by the present application;

[0095] Figure 2 is a principle diagram of the physical-information mapping in the present application;

[0096] Figure 3 is a principle diagram of the construction of the water supply network rule base in the present application;

[0097] Figure 4 This is a schematic diagram of the failure probability update logic based on Bayesian inference and hydraulic residuals in this invention. Detailed Implementation

[0098] To make the objectives, technical solutions, and advantages of this invention clearer, the embodiments of this invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0099] The evaluation method provided in this embodiment aims to solve the dual problems of traditional hydraulic models' inability to capture cascading failure processes in real time, dynamically, and stably, and the poor generalization ability and lack of physical interpretability of purely data-driven methods when samples are scarce. The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0100] The evaluation logic in this embodiment aims to address the problems of traditional hydraulic models' inability to capture cascading failures in real time and the lack of physical interpretability in purely data-driven methods. Combined with... Figure 1 As shown, a method for functional cascade evaluation of underground water supply pipelines includes the following steps:

[0101] S1. Obtain the basic topology data of the underground water supply network in the assessment area and construct a water supply network diagram model. Obtain the hydraulic data monitored in real time by sensors deployed at key nodes of the network and map it into the established water supply network diagram model to provide a basic computing platform for subsequent hydraulic state solutions.

[0102] The system imports pipeline topology data of the assessment area from a GIS geographic information system, and parses and generates a set of nodes. and pipe segment collection Composition of graphical models Simultaneously, the data interface module pulls maintenance data from the SCADA system in real time. Specifically, the SCADA (Data Acquisition and Monitoring Control) system forms a sensing link between the physical pipeline network and the digital assessment model. At the physical level, this system includes sensing devices distributed at key nodes of the pipeline network, such as pressure transmitters and electromagnetic flowmeters installed in valve wells, field control units such as RTUs or PLCs, and a communication network. The sensing devices are responsible for converting the physical state of the pipeline network, such as analog water pressure signals and flow velocity pulse signals, into digital signals and uploading them to the dispatch center server via 4G / 5G or industrial Ethernet. The assessment method described in this invention is deployed in this dispatch center, reading the aforementioned real-time monitoring data through OPC or API interfaces. ). The SCADA system not only provides real-time data flow, but also provides necessary historical sample support for the construction of the knowledge graph in subsequent step S3 and the benchmark verification of the fault fingerprint in step S5.

[0103] Specifically, step S1 specifically includes:

[0104] S11, constructing a water supply network graph model: the expression of the water supply network graph model is:

[0105] ,

[0106] Wherein: represents a set of all nodes in the network, represents a set of all pipe sections in the network.

[0107] S12, obtaining monitoring data and operation and maintenance data, the monitoring data including node monitoring pressure , pipe section monitoring flow , valve opening degree / state , pump working condition , and operation and maintenance data including work orders and repair records. In order to ensure the accuracy of the calculation, the system will automatically align the collected time series data. For missing data caused by sensor failure, a space-time interpolation algorithm is used to complete the missing data; for abnormal noise points obviously deviating from the physical range, they are rejected and assigned to the corresponding node and pipe section object in the water supply network graph model .

[0108] S2, establishing a structural failure mode library, converting the physical structural failure mode of the pipeline into an equivalent hydraulic disturbance parameter.

[0109] As shown in Figure 2 , by building a structural failure mode library in the system, the computer understands "physical damage" and maps the three typical failures in the physical world into parameters recognizable by the hydraulic model.

[0110] Specifically, the method for converting the physical structural failure mode of the pipeline into an equivalent hydraulic disturbance parameter includes:

[0111] S21, for the burst pipe or fracture mode, which belongs to the most serious failure. At the algorithm level, if the topology connection is not directly disconnected, the roughness coefficient of the pipe section is forcibly corrected to a very low value , for example, 0.001, to simulate the flow interruption caused by infinite flow resistance.

[0112] S22, for the leakage or perforation mode, which belongs to chronic failure. The system adds a virtual leakage item at the node associated with the failed pipe section , which is set as a function of the current node pressure, simulating the physical phenomenon that the higher the simulation pressure, the greater the leakage.

[0113] S23, for local buckling mode, which is common in pressure deformation of pipelines. The system modifies the effective pipe diameter of the pipe segment in the model to the reduced failure diameter , or artificially increases the local water head loss coefficient to equivalently represent the attenuation of flow capacity.

[0114] The above , and are updated as parameters to the pipe network hydraulic equation set described in step S4.

[0115] S3, construct the knowledge graph of the water supply network, input the pipe network asset properties and historical operation and maintenance records in the graph model of step S1, including pipe material, pipe diameter, pipe age, interface form and burial depth, as entities, and output the pipe segment failure prior weight, isolation boundary and secondary failure trigger rule based on the graph reasoning mechanism as the basis for step S5 prior probability calculation and step S6 cascade determination.

[0116] As Figure 3 shown is a schematic diagram of the construction principle of the rule base of the water supply network. In this embodiment, the knowledge graph technology is introduced. The system constructs a knowledge graph containing entities such as pipe segments, environment (soil, road), assets (material, pipe age) and their mutual relationships. Based on the graph, the system runs the reasoning mechanism, combines historical repair records, and automatically outputs the failure prior weight of each pipe segment under the current environment , for example, the prior weight of high-age gray cast iron pipe under heavy-load road will be automatically increased. At the same time, according to the valve distribution, the physical isolation boundary after failure is determined, and according to the hydraulic connectivity, the propagation domain rule of pressure fluctuation is determined. These priori knowledge will greatly reduce the range of subsequent search calculation.

[0117] Specifically, step S3 includes:

[0118] S31, construct the knowledge graph of the water supply network, the expression of the water supply network knowledge graph is:

[0119] ,

[0120] Wherein, the entity set includes pipe segments, nodes, valves, DMA partitions, road levels, soil environments, pipe ages, materials, failure modes, and repair resources; the relationship set includes connection relationship, supply relationship, isolation reachable relationship, same partition relationship, same repair team relationship, failure mode-disturbance type relationship.

[0121] S32. Construct a rule base based on expert knowledge base and historical fault data. By utilizing the association path search of the knowledge graph, the prior weights of the pipe segment failure modes are output. Isolation boundary range and secondary failure triggering rules, where subscript Indicates the pipe segment index. Indicates the failure mode type.

[0122] S4. Under the current failure set and valve state, introduce the equivalent hydraulic disturbance parameters generated in step S2, establish and solve the set of hydraulic equations for the pipe network consisting of the node continuity equation, the pipe segment energy equation, and the relationship equation between node pressure and head, and obtain the node pressure and pipe segment flow rate to realize the hydraulic redistribution of the pipe network; and use the pressure-driven water supply model to calculate the node service factor and the actual water supply of the node.

[0123] In this embodiment, a pressure-driven water supply model (PDA) is used, which differs from traditional models by iteratively calculating by simultaneously solving the node continuity equations and pipe segment energy equations. Specifically, step S4 includes:

[0124] S41. Establish the nodal continuity equation, the expression of which is:

[0125]

[0126] in, Indicates the current time step or iteration step of the evaluation; Indicates the current computing node. Represents nodes Adjacent nodes; Let i be the set of adjacent nodes of node i. Let the flow direction sign function be used when the flow rate is... Flow direction Add 1 if the time is right, and subtract 1 otherwise. Let k be the flow rate of the pipe segment from node i to node j. Step size Time node The user's water demand, Step size Time node The actual water supply Step size Time node The equivalent leakage amount.

[0127] S42. Establish the pipe segment energy equation based on the Hazen-Williams formula. The expression of the pipe segment energy equation is as follows:

[0128]

[0129] wherein, , denotes the step time node head, denotes the head of node denotes the length of pipe segment , denotes the Hazen-Williams coefficient of pipe segment , denotes the physical diameter of pipe segment , denotes the flow rate of pipe segment at step ; when pipe segment , according to the rules of step S2, let or or be directly disconnected, to represent different modes such as burst pipe, fracture, blockage, etc.

[0130] S43, an equation of node pressure and head is established, and the expression of the equation of node pressure and head is:

[0131]

[0132] wherein, is the node pressure, ρ is the water density, g is the gravitational acceleration, is the elevation of node ;

[0133] S44, a node service factor is calculated by using a pressure-driven water supply model, and the pressure-driven water supply model is:

[0134]

[0135] wherein, denotes the node pressure service factor, is a dimensionless coefficient with a value range of , denotes the pressure of node at step , is the minimum water supply pressure threshold, is the service pressure threshold, is the pressure-water supply nonlinear index, reflecting the nonlinear degree of the influence of pressure change on flow in the water supply network, in actual application and model construction, the value range of is usually set to 0.5 to 1.0, and the selection of the value is based on the basic principles of fluid mechanics and the characteristics of pipe materials.

[0136] Specifically, the system does not assume that the water demand of nodes is constantly satisfied, but instead uses step S44 to calculate the node service factor. When a node calculates pressure Below the minimum threshold When the service factor is 0, it means a complete water outage; when it is between Service pressure When the water supply capacity decreases in between, it is reduced according to a power function. This approach more accurately reflects the decline in water supply capacity during an accident.

[0137] S45. Calculate the actual water supply of a node based on its service factor and water demand. The calculation formula is as follows:

[0138]

[0139] in, Indicates step size Time node The actual water supply Step size Time node The user's water demand, This represents the node pressure service factor.

[0140] S5. Based on the preset monitoring nodes and monitoring pipe sections, obtain the measured data, construct the hydraulic residual as data evidence, and combine it with the prior weights output by the knowledge graph described in step S3 to calculate the posterior probability of structural failure of the candidate pipe section.

[0141] In this embodiment, the system uses real-time monitoring data to correct the model, i.e., data assimilation. For example... Figure 4 The diagram illustrates the failure probability update logic based on Bayesian inference and hydraulic residuals. The system calculates the residual vector between the measured values ​​at monitoring points and the calculated values ​​from the model in real time. The residual is then compared with a fault fingerprint generated in advance through simulation. Compare the results. If the shape of the real-time residual is similar to that of a certain pipe section... The fingerprints at the time of the tube burst were highly similar, and the prior weights shown in the knowledge graph were also similar. The probability of structural failure of the pipe segment is also relatively high; therefore, the posterior probability is calculated using the activation function. This will increase significantly. This process achieves a fusion of "data evidence" and "knowledge and experience" in the judgment.

[0142] Before performing real-time computation, this embodiment addresses the problem of insufficient actual pipe burst samples, which prevents the training of a recognition model, by pre-constructing a fault fingerprint database. The fault fingerprint... The specific generation process is as follows:

[0143] The system uses the pipeline network model constructed in step S1 as the simulation base and adopts a "traversal simulation" strategy. For each pipe segment in the pipeline network... The system sequentially assumes pipe burst failure and runs a hydraulic solver to calculate the failure status of all monitoring points in the network under these assumed failure conditions. The pressure change vector. The difference between this pressure change vector and the pressure vector under normal operating conditions is defined as the fault fingerprint of this pipe section. For example, if a pipe rupture in section A causes a pressure drop of 10m at monitoring point 1 and a pressure drop of 1m at monitoring point 2, then its fingerprint characteristics are as follows: Through the above process, the system establishes a complete fingerprint database covering all pipe segments in the entire network. In step S52, the real-time evaluation phase, the system only needs to use the residual vector captured in real-time by the SCADA system. By comparing the similarity with each fingerprint in the fingerprint database, the candidate failed pipe segment with the most matching physical characteristics can be quickly located, realizing the diagnostic logic of "supplementing data with simulation".

[0144] During the online evaluation phase, the real-time monitoring residual vector is first calculated. Pre-set fault fingerprints for each pipe section The cosine similarity is used, which characterizes the degree to which hydraulic data supports fault location; subsequently, the prior weights based on asset attribute inference output from step S3 are introduced. By merging data-driven and knowledge-driven evidence through a weighted summation method, the fault evidence strength of each candidate pipe segment is obtained. .

[0145] Specifically, step S5 includes:

[0146] S51. Construct a residual vector based on the deviation between the observed and simulated values ​​of the monitoring node set and the monitoring pipe segment set. The expression for the residual vector is:

[0147]

[0148] in, For the residual vector, For the set of monitoring nodes, For monitoring the collection of pipe sections; Representing nodes respectively Monitoring pressure and step size The pressure of time, They represent pipes Monitoring flow and step size Traffic flow at that time.

[0149] S52. Calculate the correlation between the residual vector and the preset fault fingerprint, and generate the fault evidence strength by combining the prior weights output in step S3. The formula for calculating the fault evidence strength is as follows:

[0150]

[0151] wherein, denotes the step size the failure evidence strength of pipe segment at time the theoretical residual fingerprint of pipe segment when failure occurs, generated by hydraulic simulation, the prior weight output by the knowledge graph in step S3, , is the weight coefficient, and represent the weight coefficients of different influencing factors. In order to ensure the normalization of the evaluation index, both satisfy the following constraint relationship: , and , the specific weight distribution can be adjusted according to the relative importance of each factor in the actual scene; denotes the correlation coefficient calculation function.

[0152] S53, construct a pipe segment feature vector containing pipe material, pipe age and buried depth attributes, combine the evidence strength and calculate the structure failure posterior probability through the activation function:

[0153]

[0154] wherein, denotes the step size the posterior probability of pipe segment failure at time is the Sigmoid activation function, is the pipe segment feature vector, is the attribute weight vector, is the bias term, is the evidence strength weight coefficient, and the are model parameters trained based on historical failure sample data through logistic regression or neural network algorithm.

[0155] The pipe segment feature vector in the formula is a structured numerical vector generated based on the entity attributes of the knowledge graph in step S3, which is used to quantify the inherent vulnerability of the pipe segment. Specifically, contains three-dimensional feature components:

[0156] Physical property components, including one-hot encoding of pipe type, pipe diameter normalized value, pipe age, and interface form; environmental property components, including corrosion grade of soil where the pipe segment is located, ground road load grade (e.g., main road / pedestrian path), and burial depth; operation and maintenance history components, including the cumulative number of emergency repairs in the past 5 years and the time since the last repair. In the model training stage, the parameters learn the weight relationship between the above features and the failure probability, for example, learning that "high age + grey cast iron" corresponds to a larger positive weight, so that the model can distinguish the differences in disturbance resistance of different pipe segments.

[0157] S6, calculate the overload ratio of the pipe segment according to the pipe network hydraulic redistribution result obtained in step S4, update the secondary failure probability based on the overload ratio and the secondary failure triggering rule in step S3, and fuse the secondary failure probability with the structure failure posterior probability obtained in step S5 to calculate the comprehensive failure probability of the pipe segment; arrange the comprehensive failure probability of the pipe segment at each time step in chronological order to form a comprehensive failure probability sequence.

[0158] In this embodiment, according to the current flow distribution result, the overload ratio of each pipe segment is calculated . Here, a key parameter, structure-hydraulic comprehensive capacity , is introduced. This parameter is not only a physical capacity, but also a calibrated value that comprehensively considers flow rate limitation, pipe material allowable stress, and historical experience. If the flow of a pipe segment suddenly increases , the secondary failure probability gain is calculated, and the failure probability of the pipe segment is updated. This embodiment simulates the cascading phenomenon of "one pipe burst, and the adjacent pipes break one after another due to flow overload".

[0159] Specifically, step S6 includes:

[0160] S61, calculate the overload ratio of the pipe segment according to the structure-hydraulic comprehensive capacity preset for the water supply network and the pipe segment flow output in step S4, and the calculation formula is:

[0161]

[0162] wherein, is the overload ratio, i.e., the utilization rate of the structure performance; is the structure-hydraulic comprehensive capacity of the pipe segment, which is a comprehensive index reflecting the maximum bearing capacity of the pipe segment. The calculation of this value comprehensively considers the design flow of the pipe segment, the structural strength limit of the pipe material, and the historical maximum load data, and is obtained by weighted comprehensive calculation; represents the flow of the pipe segment at step .

[0163] S62, calculating the secondary failure probability gain based on the overload ratio calculated in step S61, the calculation formula is:

[0164]

[0165] wherein, is the secondary failure probability gain, is the overload sensitivity coefficient, is an empirical parameter, which is used to quantify the response sensitivity of the system to the flow overload condition, and its specific value is not fixed, but is obtained through regression analysis and calibration of the historical operation data of the regional water supply network;

[0166] and the structure failure posterior probability obtained in step S5 is fused with the secondary failure probability gain obtained in this step according to the following formula: to calculate the comprehensive failure probability:

[0167]

[0168] wherein, is the comprehensive failure probability at the current time after fusing the data-driven structure failure evidence and the model-driven hydraulic overload risk; if the next time step evaluation is needed, let .

[0169] S7, introducing a cascade consistency constraint, monotonically projecting the comprehensive failure probability sequence obtained in step S6 to meet the monotone increasing nature of the failure set, and constructing a system function index sequence based on the actual water supply quantity of the node calculated in step S4, and imposing a monotone regression constraint on the system function index sequence to eliminate the non-physical function rebound.

[0170] In this embodiment, in order to prevent the phenomenon of violating the physical law caused by numerical calculation error, a strong constraint mechanism is introduced:

[0171] Monotonicity projection: for the failure set , the system is forced to monotonically increase with time, that is, the pipe segment that has been determined to fail will not automatically recover unless there is a manual intervention signal.

[0172] Function index regression: the system calculates the system function index , that is, the full network weighted water supply satisfaction rate. Without implementing the fixed strategy of repair, the system is forced to constrain to appear non-physical rebound.

[0173] ​Causal consistency: for any pipe segment newly determined to be failed, the system will check its trigger condition retrospectively. Only when one of the three conditions, i.e. overload trigger, knowledge rule trigger (e.g. located in high risk area) or data evidence trigger (residual error match) is met, the secondary failure is confirmed to be valid. This effectively avoids false positives caused by algorithm hallucination.

[0174] In particular, step S7 comprises:

[0175] S71, the comprehensive failure probability sequence obtained in step S6 is projected monotonically, and the expression of the monotonically projected failure probability sequence is:

[0176]

[0177] wherein, is the final failure determination probability after monotonous correction, is the historical time step index, and the threshold is a failure set is generated , to ensure that the failure set satisfies , is a critical threshold for determining whether a node is in a failure state, which is an adjustable parameter, and its value range is usually set between 0.5 and 0.8 according to different requirements for system reliability. When the calculated state value is lower than the threshold, the node is determined to be failed.

[0178] S72, the system function index is constructed using the actual water supply amount of the node obtained in step S4, and the calculation formula is:

[0179]

[0180] wherein, is a set of user nodes with water demand requirements, is the importance weight of the node , which is set according to factors including but not limited to: the type of user in the area where the node is located, such as the weight of key guarantee units like hospitals and schools, the population density of the area, and the topological importance of the node, such as whether it is a hub node; represents the system function satisfaction rate calculated based on the current hydraulic state only; is the actual water supply amount of the node at step ; is the water demand of the user of the node at step ; represents the node pressure service factor; then, a monotonous regression constraint is applied to obtain the corrected effective system function index , and ​Monotone regression constraint ensures that system performance index is monotone non-increasing with the cascading process, eliminating the non-physical function rebound caused by numerical fluctuations.

[0181] S73, perform structure-function causal consistency check to ensure that the secondary failure pipe segment meets at least one of the following conditions: overload trigger, knowledge rule trigger, or data evidence trigger:

[0182] When a pipe segment is determined to be a secondary failure , at least one trigger condition must be met:

[0183] Overload trigger, i.e. ;

[0184] Or knowledge rule trigger, i.e. Give trigger labels such as "isolation unreachable / low pressure persistence / material high fragility";

[0185] Or data evidence trigger, the residual fingerprint similarity corresponding to the pipe segment exceeds the threshold.

[0186] S8, calculate the functional cascading comprehensive loss index based on the system function index sequence after step S7 consistency constraint, and quantify the marginal contribution of each pipe segment to the functional cascading comprehensive loss index, and identify the key weak link and high-risk cascading path of the pipe network.

[0187] In this embodiment, the system function index in the entire cascading process is integrated (weighted sum) to obtain the functional cascading comprehensive loss index . The lower the index, the greater the cumulative social impact caused by cascading failure. Further, the system uses "one by one elimination method" or "sensitivity analysis method" to calculate the contribution of each pipe segment to the total loss . Finally, the system highlights the pipe segment with the highest contribution (key weak link) on the human-computer interaction interface and dynamically displays the high-risk cascading propagation path (for example: pipe segment A burst causes pipe segment B overload leading to regional C pressure loss), providing accurate decision basis for repair personnel. The system summarizes the whole process data.

[0188] Specifically, step S8 includes:

[0189] S81, calculate the functional cascading comprehensive loss index based on the system function index after step S7 consistency constraint and event weight, and the calculation formula of the functional cascading comprehensive loss index is:

[0190]

[0191] Wherein, is the functional cascading comprehensive loss index, Step size Time weight when the step size isIn the embodiment, exponential decay is used for assignment, that is, data closer to the current time is given a greater weight to better reflect the real-time state change of the system, Total time length step number of cascade evaluation, System function index after consistency constraint;

[0192] S82, based on the cascade failure under different failure sets, the weak link contribution degree of each pipe section is calculated to identify the key weak link and high-risk cascade path, and the calculation formula of the weak link contribution degree is:

[0193]

[0194] Wherein, Weak link contribution degree, Loss obtained by cascade evaluation with failure set F, The loss of the pipe section After the pipe section is removed from the failure set, the counterfactual evaluation loss is arranged in descending order according to the numerical value to identify the key weak link.

[0195] Through the above steps, the problem of oscillation of cascade evaluation results in the prior art is effectively solved, and a scientific basis is provided for emergency scheduling and weak point management of the water supply network.

[0196] The present application avoids the common non-physical oscillation in cascade evaluation, that is, failure rollback and service recovery, through the monotonicity of failure set and the monotonicity of function degradation, so that the results can be reviewed and compared; the present application explicitly includes structural performance failure mode, isolation boundary and hydraulic disturbance into knowledge graph and rule base, and secondary failure must be given triggering basis to realize consistency of "evidence-mechanism-result"; the knowledge prior in the present application provides transferable structural vulnerability and isolation boundary information, which can maintain available cascade diagnosis ability when the monitoring points are sparse or the historical samples are insufficient, combined with monitoring residual assimilation; the present application outputs the core weak link ranking with comprehensive loss And contribution degree , which directly interfaces with engineering decisions such as "priority update / priority inspection / priority reinforcement / priority partition reconstruction".

[0197] The underground water supply pipeline function cascade evaluation method described in the present application can be realized by an electronic device. The electronic device comprises at least one processor (CPU), a memory, an input / output interface and a communication module. The memory can be a volatile memory (such as RAM), a non-volatile memory (such as ROM, flash memory) or a combination of the two, which stores a computer program. The computer program contains executable instructions which, when executed by the processor, control the electronic device to perform all or part of the operations of steps S1 to S8 described above, thereby realizing the evaluation method of the present application. The electronic device can be a server, a workstation, or a computing node integrated into an existing SCADA system of a water supply dispatch center. The input / output interface is used to receive data from external systems such as SCADA, GIS, etc., and can output evaluation results (such as a list of weak links, a risk path diagram) to a display device or other decision support systems.

[0198] The present embodiment has been tested in a real DMA pipe network through the above steps. When the SCADA system alarms an abnormal pressure in a certain area, the system automatically completes the entire process from data reception to report generation within 3 minutes, accurately locates the initial pipe burst position, predicts 2 pipe sections at risk due to overload within the next 30 minutes, and provides a priority maintenance list containing 5 key weak pipe sections. The evaluation results are verified by on-site repair personnel and serve as an important basis for the reconstruction planning of the pipe network in the area.

[0199] The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structural transformation made under the inventive concept of the present application, using the contents of the present application specification and drawings, or direct / indirect application in other related technical fields is included in the patent protection scope of the present application.

Claims

1. A method for functional cascade assessment of underground water supply pipelines, characterized in that, The method comprises the following steps: S1, acquiring basic topological data of an evaluation area underground water supply pipe network and constructing a water supply pipe network graph model, acquiring real-time monitored hydraulic data, and mapping the hydraulic data to the water supply pipe network graph model; S2, establishing a structure failure mode library, and converting physical structure failure modes of the pipe into equivalent hydraulic disturbance parameters; S3, constructing a water supply pipe network knowledge graph, taking pipe network asset properties and historical operation and maintenance records in the graph model of step S1, including pipe material, pipe diameter, pipe age, interface form, and buried depth, as entity inputs, and outputting pipe segment failure prior weights, isolation boundaries, and secondary failure triggering rules based on a graph reasoning mechanism; S4, under the current failure set and valve state, introducing the equivalent hydraulic disturbance parameters generated in step S2, establishing a pipe network hydraulic equation set composed of a node continuity equation, a pipe segment energy equation, and a node pressure and head relationship equation, and solving to obtain node pressure and pipe segment flow, thereby realizing pipe network hydraulic redistribution; and using a pressure-driven water supply model to calculate node service factors and node actual water supply amounts; S5, acquiring measured data based on preset monitoring nodes and monitoring pipe segments, constructing hydraulic residuals as data evidence, and combining the prior weights output by the knowledge graph in step S3 to calculate structure failure posterior probabilities of candidate pipe segments; S6, calculating pipe segment overload ratios based on the pipe network hydraulic redistribution results obtained in step S4, updating secondary failure probabilities based on the overload ratios and the secondary failure triggering rules in step S3, and fusing the secondary failure probabilities and the structure failure posterior probabilities obtained in step S5 to calculate pipe segment comprehensive failure probabilities; arranging the pipe segment comprehensive failure probabilities of each time step in chronological order to form a comprehensive failure probability sequence; S7, introducing a cascading consistency constraint, performing a monotone projection on the comprehensive failure probability sequence obtained in step S6 to meet the monotone increasing nature of the failure set, and constructing a system function index sequence based on the node actual water supply amounts calculated in step S4, and applying a monotone regression constraint to the system function index sequence to eliminate non-physical function recovery; S8, calculating a functional cascading comprehensive loss index based on the system function index sequence after the consistency constraint in step S7, and quantifying marginal contributions of each pipe segment to the functional cascading comprehensive loss index, thereby identifying key weak links and high-risk cascading paths of the pipe network.

2. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, Step S1 specifically comprises: S11, constructing a water supply pipe network graph model: the expression of the water supply pipe network graph model is: ， wherein: denotes the set of all nodes in the pipe network, denotes the set of all pipe segments in the pipe network; S12, acquire monitoring data and operation and maintenance data: the monitoring data includes node monitoring pressure , pipe section monitoring flow , valve opening degree / state , pump working condition ; the operation and maintenance data includes work order and repair record; after uniform timestamp, missing value completion and abnormal value elimination of the above data, the data is assigned to the corresponding node and pipe section object in the water supply network graph model .

3. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, In step S2, the method for converting the physical structure failure modes of the pipe into equivalent hydraulic disturbance parameters comprises: S21, for a burst or break mode, the roughness coefficient of the failed pipe segment is replaced by a failure state coefficient S21, for a burst or break mode, the roughness coefficient of the failed pipe segment is replaced by a failure state coefficient or the pipe segment is disconnected in the topology; S22, for the leakage or perforation mode, introduce a pressure-dependent leakage term at the end nodes of the failed pipe segment ; S23, for local buckling mode, the effective pipe diameter of the pipe section is modified to the reduced failure diameter or the local head loss coefficient is increased or the local head loss coefficient is increased The , and are updated as parameters into the pipe network hydraulic equation set described in step S4.

4. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, Step S3 specifically comprises: S31, constructing a water supply pipe network knowledge graph: the expression of the water supply pipe network knowledge graph is: ， Wherein, the entity set Including pipe segment, node, valve, DMA partition, road level, soil environment, pipe age, material, failure mode, repair resource; relationship set Including connection relationship, supply relationship, isolation reachable relationship, same partition relationship, same repair team relationship, failure mode-disturbance type relationship; S32, constructing a rule library based on the expert knowledge base and historical failure data , using the correlation path search of the knowledge graph to output the prior weight of the pipe segment failure mode , the isolation boundary range and the secondary failure triggering rule, wherein the subscript represents the pipe segment index, represents the failure mode type.

5. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, Step S4 specifically comprises: S41, establishing a node continuity equation: the expression of the node continuity equation is: in, Indicates the current time step or iteration step of the evaluation; Indicates the current computing node. Represents nodes Adjacent nodes; Let i be the set of adjacent nodes of node i. For the flow sign function, when the flow rate is from Flow direction Add 1 if the time is right, and subtract 1 otherwise. Let k be the flow rate of the pipe segment from node i to node j. Step size Time node The user's water demand, Step size Time node The actual water supply Step size Time node The equivalent leakage amount; S42, establishing a pipe segment energy equation based on a Hazen-Williams formula: the expression of the pipe segment energy equation is: wherein, represents the step the water head height at the node at the step the water head height at the node at the step the water head height at the node represents the length of the pipe section at the step represents the length of the pipe section at the step represents the physical diameter of the pipe section at the step represents the flow rate of the pipe section at the step at the step at the step or or is directly disconnected to represent the mode of burst pipe, fracture, blockage; S43, establishing a node pressure and head relationship equation: the expression of the node pressure and head relationship equation is: wherein, Pn is the nodal pressure, p is the water density, g is the gravitational acceleration, Pn is the nodal pressure, p is the water density, g is the gravitational acceleration, the elevation of the node S44, calculating node service factors using a pressure-driven water supply model: the pressure-driven water supply model is: wherein, represents the node pressure service factor, is a dimensionless coefficient with a value ranging from represents the step size the pressure of the node is the minimum water supply pressure threshold, is the service pressure threshold, is the pressure-water supply nonlinearity index;​​ S45, calculating the actual water supply of the node based on the node service factor and the water demand, and the calculation formula is: wherein, representing a step size a time node actual water supply, a step size a time node user water demand, representing a node pressure service factor.

6. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, Step S5 specifically includes: S51, constructing a residual vector based on the deviation between the observation value and the simulation value of the monitoring node set and the monitoring pipe section set, and the expression of the residual vector is: in, For the residual vector, For the set of monitoring nodes, For monitoring the collection of pipe sections; Representing nodes respectively Monitoring pressure and step size The pressure of time, They represent pipes Monitoring flow and step size Traffic flow at that time; S52, calculating the correlation of the residual vector and the preset fault fingerprint, and generating the fault evidence strength in combination with the prior weight output in step S3, and the calculation formula of the fault evidence strength is: in, Indicates step size Time Management Section The strength of evidence of failure For pipe section The theoretical residual fingerprint at the time of failure is generated by hydraulic simulation. The prior weights are the output of the knowledge graph in step S3. , These are the weighting coefficients; This represents the function for calculating the correlation coefficient. S53, constructing a pipe section feature vector containing pipe material, pipe age, and buried depth attributes, combining the evidence strength, and calculating the structure failure posterior probability through an activation function, and the calculation formula of the structure failure posterior probability is: wherein, step size time pipe segment posterior probability of structural failure occurring, sigmoid activation function, pipe segment feature vector, attribute weight vector, bias term, evidence strength weight coefficient, the all are model parameters obtained by training through a logistic regression or neural network algorithm based on historical failure sample data.

7. The functional cascade assessment method for underground water supply pipes according to claim 1, characterized by, Step S6 specifically includes: S61, calculating the pipe section overload ratio according to the preset structure-hydraulic comprehensive capacity of the water supply network and the pipe section flow output in step S4, and the calculation formula is: wherein, is an overload ratio, i.e. a structural performance utilization, is a pipe section structure-hydraulic integrated capacity; denotes a step size denotes a flow rate of the pipe section at the time. S62, calculating the secondary failure probability gain based on the overload ratio in step S61, and the calculation formula is: wherein, is the secondary failure probability gain, is the overload sensitivity coefficient; And according to the following formula, the posterior probability of structural failure obtained in step S5 is... The gain on the secondary failure probability obtained in this step Perform fusion and calculate the overall failure probability: where, is the current time step integrated failure probability after fusing data-driven structural failure evidence with model-driven hydraulic overload risk; if next time step evaluation is required, then let .

8. The functional cascade assessment method for underground water supply pipes according to claim 7, characterized by, Step S7 specifically includes: S71. The comprehensive failure probability sequence obtained in step S6 Perform monotonic projection, the expression for which is: wherein, wherein, is the final failure decision probability after monotonic correction, is the history time step index, and generates the failure set ensures that the failure set satisfies ; S72, constructing a system function index using the actual water supply of the node obtained in step S4, and the calculation formula is: wherein, is a set of user nodes with water demand requirements, is a node importance weight; represents the system function satisfaction rate calculated only based on the current hydraulic state; is the actual water supply amount of the node at step ; is the user water demand of the node at step ; represents the node pressure service factor; subsequently, a monotonic regression constraint is imposed thereon to obtain a modified effective system function index , set , the monotonic regression constraint ensures that the system performance index is monotonically non-increasing with the cascade process; S73, performing structure-function causal consistency checking to ensure that the secondary failure pipe section meets at least one of the following conditions: overload triggering, knowledge rule triggering, or data evidence triggering: When a pipe section is determined to be a secondary failure, i.e. at least one trigger condition must be met, the trigger condition comprising an overload trigger, a knowledge rule trigger or a data evidence trigger; In the formula, the determination condition of overload triggering is: ; The determination condition of the knowledge rule triggering is that the rule base gives a trigger label of isolation unreachable, low pressure persistence, or material high fragility; The determination condition of the data evidence triggering is that the similarity of the residual fingerprint corresponding to the pipe section exceeds a threshold.

9. The functional cascade assessment method for a water-supply pipe under the ground according to Claim 1, wherein Step S8 specifically includes: S81, calculating a functional cascade comprehensive loss index based on the system function index and the event weight after the consistency constraint in step S7, and the calculation formula of the functional cascade comprehensive loss index is: wherein, is a functional cascade comprehensive loss index, is a time weight, is a total duration step of cascade evaluation, is a system function index after consistency constraint; S82, calculating the weak link contribution degree of each pipe section based on the cascade failure under different failure sets to identify the key weak link and the high-risk cascade path, and the calculation formula of the weak link contribution degree is: wherein, contribution of weak links, represents the loss obtained by cascading evaluation with failure set F, is the loss of the pipe segment the counterfactual evaluation loss after removing the failure set from the loss, ranked in numerical descending order to identify the key weak links.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and capable of running on the processor, characterized in that, The processor executes the computer program to implement the underground water supply pipeline function cascade evaluation method according to any one of claims 1-9.

Images